Cloud Memory Manage Tool
memory_manageStore and retrieve team-scoped semantic notes across conversations. Supports search, add, delete, and bulk upload to remember preferences, constraints, and decisions.
Instructions
Team-scoped semantic memory — short notes agents store and retrieve across conversations (preferences, constraints, prior decisions). Backed by pgvector with HNSW indexing; cross-team isolation enforced by TeamScope. Distinct from knowledge_manage: this stores unstructured short notes (1-2 paragraphs), knowledge_manage stores document corpora.
When to use: an agent needs to recall a stable fact about the user/team across runs, or wants to remember a decision so the next conversation starts with that context. Do NOT use for transient run-state — use the experiment artifact system instead.
Actions:
search (read; semantic) — query (string), optional limit (default 20, max 100). Returns ranked entries with similarity score 0..1; entries below threshold 0.6 are filtered.
list_recent (read) — optional limit (default 50). Reverse-chronological by created_at.
stats (read) — returns { total_entries, embedding_dim, last_write_at, storage_bytes }.
add (write — costs ~1 credit per ~1k tokens for embedding) — content (string, max 4000 chars), optional metadata (object).
delete (DESTRUCTIVE) — memory_id (UUID). Hard delete; not recoverable.
upload_knowledge (write — costs embedding credits, async) — one of: file content (string), url. Bulk-ingests into memory; returns job_id. Up to 50 MB per call.
Cloud note: supabase_provision not available in cloud (provisioning happens at platform level). Errors: 401, 403, 404 if memory_id missing, 413 on oversized payload, 422 validation, 429 rate limit, 503 if budget exhausted.
Input Schema
| Name | Required | Description | Default |
|---|---|---|---|
| action | Yes | Action to perform: search, list_recent, stats, add, delete, upload_knowledge | |
| deadline_ms | No | Optional: max wall-clock time (ms) the tool may spend. If exceeded during the call, returns a DEADLINE_EXCEEDED error. Minimum 100 ms. Leave unset for no deadline. | |
| query | Yes | Search keyword to match against memory content | |
| agent_id | No | Filter by agent UUID | |
| limit | No | Max results to return (default 10, max 100) | |
| min_confidence | No | Minimum confidence score to filter results (0.0–1.0, default 0.0 to include all) | |
| category | No | Filter by memory category: preference, knowledge, context, behavior, goal | |
| search_mode | No | Retrieval mode: semantic=flat keyword search, local=1-hop graph traversal from matched entities, global=high-centrality entities, hybrid=semantic+local merged, mix=semantic+global merged | semantic |
| tags | No | Filter by tags — only return memories containing ANY of these tags. E.g. ["barsy:client", "barsy:shared"]. Omit to return all memories regardless of tags. | |
| topic | No | Namespace pre-filter by topic slug, e.g. "auth_migration". Narrows the search to a named context before the vector scan for higher precision. | |
| source_type | No | Filter by source type (e.g. execution, manual, signal) | |
| content | Yes | The memory text to store | |
| project_id | No | Associate this memory with a specific project UUID (optional) | |
| confidence | No | Confidence score 0.0–1.0. Default: 1.0 for manually added memories | |
| metadata | No | Additional structured metadata (key-value pairs) | |
| memory_ids | Yes | Array of memory UUIDs to delete |